Removing a dimension or metric from your dataset's field configuration stops that field from being collected going forward — but it does not erase data you've already ingested. Here's why the field can still appear in exploration, widget previews, and the dimension/metric picker, and why that's expected.

Overview

When you set up a dataset, you choose which dimensions and metrics to bring in. Later, you can remove a field from the dataset's field customization/configuration if you no longer want to collect it.

Removing a field is allowed and takes effect immediately for future data collection. However, you may notice the field still shows up when you explore your data or build a widget. This is expected behavior. This article explains what removal actually does and why previously-collected data for that field remains visible.

What removing a field actually does

Removing a dimension or metric from a dataset's configuration means one thing:

"Stop bringing this field in when data is ingested going forward."

That's it. From the moment you remove the field, no new data for it will be collected on future syncs.

What removal does not do:

• It does not delete the data that was already collected for that field.
• It does not remove the field from the dataset "at rest."

Because earlier ingestion already brought that field in and saved it, the dataset still contains the field and its historical values. We intentionally preserve this historical data — it cannot be, and is not, removed when you drop the field from the configuration.

Why you still see the field

Data exploration, widget previews, and the dimension/metric picker all show the fields that exist in the dataset's saved schema — that is, everything the dataset has collected, including your historical data. They do not show only the fields currently being collected.

So after you remove a field, expect the following for the historical date range where data already exists:

• Data exploration will still show the column.
• Template Editor previews — for both new and existing widgets built on that dataset — will still show the field.
• The dimension/metric picker will still list the field, so it can still be selected.

This can feel like a disconnection between the configuration (where the field is gone) and the editor/exploration (where it still appears), but both are behaving correctly. The field is gone from future collection; its historical data is still there to be viewed and reported on.

What this means / FAQ

Did my removal fail? I still see the field.
No. The removal succeeded — the field is no longer part of go-forward ingestion. What you're seeing is the historical data that was already collected before you removed the field. That data is preserved by design.

Will new data keep coming in for this field?
No. After removal, no new values for that field will be collected on future syncs. Only the previously-collected (historical) values remain.

Why can I still pick the removed field in a widget?
The dimension/metric picker reflects the dataset's saved schema, which includes historical data. As long as historical values exist for the field, it remains selectable and will render for the date range where that data exists.

What happens to a widget that uses the removed field?
It will continue to display data for the historical range where values were collected. For dates after removal, no new values will be available for that field.

How do I make the field stop appearing everywhere?
Removing the field only stops future collection; it does not delete historical data, so the field remains visible for the range that already has data. If you need the field to no longer appear at all, the historical data itself would need to be addressed — reach out to NinjaCat Support if that's your goal.